Calculation of the lake evaporation may be done using artificial neural network method and weather parameters are needed to estimate. You may see some of the prediction models of ANN.

Thanks Parameshwar for your timely answer and the file.

Thanks Parameshwar for your timely answer and the file.

This is a really hot topic. My experience is that the Bulk Transfer Model may be a good choice if you have meausred E over lake to calibrtate the transfer coefficients which highly depend on not only the zom, zov but also the atmospheric stability. Such method is able to calculate at a high temporal resolution (e.g. 30mins). However, if you do not have such measured data, the issue turns out to be more complicated. The main issue is not about which model is best but depends on how's accurancy of your forcing data. For example, the wind speed. You may be able to get wind data easyily from the stations, but you need to realize that the lake-breeze would greatly impact the wind data which were observed in lake bank. Another key issue is how to calculate the lake water storage. Unlike the soil heat flux which is easy to meausred by the HFP01SC (or something like that), the lake water heat storage usually show obvious time lag when compare with the solar radiation. You may read some paper that discuss the convetional methods (e.g. Turc, Makkink, Thornthwaite etc) that calculate the lake evaporation by only using the T and Rs, but these methods obviously do not take the water heat storage into consideration. You may also found some studies used the famous Penman or the FAO'S PM modls, but they usually implictly show how they cope with the G was calculated in these models. I also would like to remind the these methods depend highly on the accurancy of the wind data, which may contains large uncertanities if we use the measured wind from the adjacent meteorological stations.

Hope it is helpful. 

Ma Ning, thanks so much for your professional answer. We can discuss this issue further. I have two flux tower's 4 year data, am working on this issue. If you have any idea, pls contact me directly.

From my page you can find my some paper on this issue.

In my view there are two hot topic which may be significant to our understanding the ET, one is the partition ET into T and E, you may read a couple of paper in NS in recently years, especially the Jasechko et al. 2013. How the vegetation variation impact the T/ET  and water use efficiency in different climatic regions? I believe this is very interesting topic. You may read Lixin's paper Global synthesis of vegetation control on evapotranspiration partitioning to do a case study in your sites.

Another interesting issue is that how the water and energy interactively influence the energy partitioning and ET variation. Mostly the water may be important for most grassland ecosystem. But the research in Mediterranean grassland (you know the rainy season is in winter for Mediterranean climatic region. e.g, California) show that the available energy also plays important role in determining the ET. For this issue you may read: "Contrasting precipitation seasonality influences evapotranspiration dynamics in water-limited shrublands" OR "Interannual variability of evapotranspiration and energy exchange over an annual grassland in California".

You can also email me  for more discussions.

Hope it is helpful

Ning